/[pcre]/code/trunk/Tech.Notes
ViewVC logotype

Contents of /code/trunk/Tech.Notes

Parent Directory Parent Directory | Revision Log Revision Log


Revision 23 - (hide annotations) (download)
Sat Feb 24 21:38:41 2007 UTC (7 years, 8 months ago) by nigel
File size: 9678 byte(s)
Load pcre-2.00 into code/trunk.

1 nigel 3 Technical Notes about PCRE
2     --------------------------
3    
4     Many years ago I implemented some regular expression functions to an algorithm
5     suggested by Martin Richards. These were not Unix-like in form, and were quite
6     restricted in what they could do by comparison with Perl. The interesting part
7     about the algorithm was that the amount of space required to hold the compiled
8     form of an expression was known in advance. The code to apply an expression did
9     not operate by backtracking, as the Henry Spencer and Perl code does, but
10     instead checked all possibilities simultaneously by keeping a list of current
11     states and checking all of them as it advanced through the subject string. (In
12     the terminology of Jeffrey Friedl's book, it was a "DFA algorithm".) When the
13     pattern was all used up, all remaining states were possible matches, and the
14     one matching the longest subset of the subject string was chosen. This did not
15     necessarily maximize the individual wild portions of the pattern, as is
16     expected in Unix and Perl-style regular expressions.
17    
18     By contrast, the code originally written by Henry Spencer and subsequently
19     heavily modified for Perl actually compiles the expression twice: once in a
20     dummy mode in order to find out how much store will be needed, and then for
21 nigel 23 real. The execution function operates by backtracking and maximizing (or,
22     optionally, minimizing in Perl) the amount of the subject that matches
23     individual wild portions of the pattern. This is an "NFA algorithm" in Friedl's
24     terminology.
25 nigel 3
26     For this set of functions, I tried at first to invent an algorithm that used an
27     amount of store bounded by a multiple of the number of characters in the
28     pattern, to save on compiling time. However, because of the greater complexity
29     in Perl regular expressions, I couldn't do this. In any case, a first pass
30     through the pattern is needed, in order to find internal flag settings like
31 nigel 23 (?i) at top level. So it works by running a very degenerate first pass to
32     calculate a maximum store size, and then a second pass to do the real compile -
33     which may use a bit less than the predicted amount of store. The idea is that
34     this is going to turn out faster because the first pass is degenerate and the
35     second can just store stuff straight into the vector. It does make the
36     compiling functions bigger, of course, but they have got quite big anyway to
37     handle all the Perl stuff.
38 nigel 3
39     The compiled form of a pattern is a vector of bytes, containing items of
40     variable length. The first byte in an item is an opcode, and the length of the
41     item is either implicit in the opcode or contained in the data bytes which
42     follow it. A list of all the opcodes follows:
43    
44     Opcodes with no following data
45     ------------------------------
46    
47     These items are all just one byte long
48    
49     OP_END end of pattern
50     OP_ANY match any character
51     OP_SOD match start of data: \A
52     OP_CIRC ^ (start of data, or after \n in multiline)
53     OP_NOT_WORD_BOUNDARY \W
54     OP_WORD_BOUNDARY \w
55     OP_NOT_DIGIT \D
56     OP_DIGIT \d
57     OP_NOT_WHITESPACE \S
58     OP_WHITESPACE \s
59     OP_NOT_WORDCHAR \W
60     OP_WORDCHAR \w
61 nigel 23 OP_EODN match end of data or \n at end: \Z
62     OP_EOD match end of data: \z
63 nigel 3 OP_DOLL $ (end of data, or before \n in multiline)
64    
65    
66     Repeating single characters
67     ---------------------------
68    
69     The common repeats (*, +, ?) when applied to a single character appear as
70     two-byte items using the following opcodes:
71    
72     OP_STAR
73     OP_MINSTAR
74     OP_PLUS
75     OP_MINPLUS
76     OP_QUERY
77     OP_MINQUERY
78    
79     Those with "MIN" in their name are the minimizing versions. Each is followed by
80     the character that is to be repeated. Other repeats make use of
81    
82     OP_UPTO
83     OP_MINUPTO
84     OP_EXACT
85    
86     which are followed by a two-byte count (most significant first) and the
87     repeated character. OP_UPTO matches from 0 to the given number. A repeat with a
88     non-zero minimum and a fixed maximum is coded as an OP_EXACT followed by an
89     OP_UPTO (or OP_MINUPTO).
90    
91    
92     Repeating character types
93     -------------------------
94    
95     Repeats of things like \d are done exactly as for single characters, except
96     that instead of a character, the opcode for the type is stored in the data
97     byte. The opcodes are:
98    
99     OP_TYPESTAR
100     OP_TYPEMINSTAR
101     OP_TYPEPLUS
102     OP_TYPEMINPLUS
103     OP_TYPEQUERY
104     OP_TYPEMINQUERY
105     OP_TYPEUPTO
106     OP_TYPEMINUPTO
107     OP_TYPEEXACT
108    
109    
110     Matching a character string
111     ---------------------------
112    
113     The OP_CHARS opcode is followed by a one-byte count and then that number of
114     characters. If there are more than 255 characters in sequence, successive
115     instances of OP_CHARS are used.
116    
117    
118     Character classes
119     -----------------
120    
121 nigel 13 OP_CLASS is used for a character class, and OP_NEGCLASS for a negated character
122     class, provided there are at least two characters in the class. If there is
123     only one character, OP_CHARS is used for a positive class, and OP_NOT for a
124     negative one. A set of repeating opcodes (OP_NOTSTAR etc.) are used for a
125     repeated, negated, single-character class.
126 nigel 3
127 nigel 13 Both OP_CLASS and OP_NEGCLASS are followed by a 32-byte bit map containing a 1
128     bit for every character that is acceptable. The bits are counted from the least
129     significant end of each byte. The reason for having two opcodes is to cope with
130     negated character classes when caseless matching is specified at run time but
131     not at compile time. If it is specified at compile time, the bit map is built
132     appropriately. This is the only time that a distinction is made between
133     OP_CLASS and OP_NEGCLASS, when the bit map was built in a caseful manner but
134     matching must be caseless. For OP_CLASS, a character matches if either of its
135     cases is in the bit map, but for OP_NEGCLASS, both of them must be present.
136 nigel 3
137 nigel 13
138 nigel 3 Back references
139     ---------------
140    
141     OP_REF is followed by a single byte containing the reference number.
142    
143    
144     Repeating character classes and back references
145     -----------------------------------------------
146    
147     In both cases, the repeat information follows the base item. The matching code
148     looks at the following opcode to see if it is one of
149    
150     OP_CRSTAR
151     OP_CRMINSTAR
152     OP_CRPLUS
153     OP_CRMINPLUS
154     OP_CRQUERY
155     OP_CRMINQUERY
156     OP_CRRANGE
157     OP_CRMINRANGE
158    
159     All but the last two are just single-byte items. The others are followed by
160     four bytes of data, comprising the minimum and maximum repeat counts.
161    
162    
163     Brackets and alternation
164     ------------------------
165    
166     A pair of non-identifying (round) brackets is wrapped round each expression at
167     compile time, so alternation always happens in the context of brackets.
168     Non-identifying brackets use the opcode OP_BRA, while identifying brackets use
169     OP_BRA+1, OP_BRA+2, etc. [Note for North Americans: "bracket" to some English
170     speakers, including myself, can be round, square, or curly. Hence this usage.]
171    
172     A bracket opcode is followed by two bytes which give the offset to the next
173     alternative OP_ALT or, if there aren't any branches, to the matching KET
174     opcode. Each OP_ALT is followed by two bytes giving the offset to the next one,
175     or to the KET opcode.
176    
177     OP_KET is used for subpatterns that do not repeat indefinitely, while
178     OP_KETRMIN and OP_KETRMAX are used for indefinite repetitions, minimally or
179     maximally respectively. All three are followed by two bytes giving (as a
180     positive number) the offset back to the matching BRA opcode.
181    
182     If a subpattern is quantified such that it is permitted to match zero times, it
183     is preceded by one of OP_BRAZERO or OP_BRAMINZERO. These are single-byte
184     opcodes which tell the matcher that skipping this subpattern entirely is a
185     valid branch.
186    
187     A subpattern with an indefinite maximum repetition is replicated in the
188     compiled data its minimum number of times (or once with a BRAZERO if the
189     minimum is zero), with the final copy terminating with a KETRMIN or KETRMAX as
190     appropriate.
191    
192     A subpattern with a bounded maximum repetition is replicated up to the maximum
193     number of times, with BRAZERO or BRAMINZERO before each replication after the
194     minimum. In effect, (abc){2,5} becomes (abc)(abc)(abc)?(abc)?(abc)?.
195    
196    
197     Assertions
198     ----------
199    
200 nigel 23 Forward assertions are just like other subpatterns, but starting with one of
201     the opcodes OP_ASSERT or OP_ASSERT_NOT. Backward assertions use the opcodes
202     OP_ASSERTBACK and OP_ASSERTBACK_NOT, and the first opcode inside the assertion
203     is OP_REVERSE, followed by a two byte count of the number of characters to move
204     back. A separate count is present in each alternative of a lookbehind
205     assertion, allowing them to have different fixed lengths.
206 nigel 3
207    
208     Once-only subpatterns
209     ---------------------
210    
211     These are also just like other subpatterns, but they start with the opcode
212     OP_ONCE.
213    
214    
215 nigel 23 Conditional subpatterns
216     -----------------------
217    
218     These are like other subpatterns, but they start with the opcode OP_COND. If
219     the condition is a back reference, this is stored at the start of the
220     subpattern using the opcode OP_CREF followed by one byte containing the
221     reference number. Otherwise, a conditional subpattern will always start with
222     one of the assertions.
223    
224    
225     Changing options
226     ----------------
227    
228     If any of the /i, /m, or /s options are changed within a parenthesized group,
229     an OP_OPT opcode is compiled, followed by one byte containing the new settings
230     of these flags. If there are several alternatives in a group, there is an
231     occurrence of OP_OPT at the start of all those following the first options
232     change, to set appropriate options for the start of the alternative.
233     Immediately after the end of the group there is another such item to reset the
234     flags to their previous values. Other changes of flag within the pattern can be
235     handled entirely at compile time, and so do not cause anything to be put into
236     the compiled data.
237    
238    
239 nigel 3 Philip Hazel
240 nigel 23 September 1998

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12